Study on the origin traceability of Menthae Haplocalycis Herba in China based on multivariate data fusion combined with an artificial intelligence classification algorithm

In this study, characteristics of Menthae Haplocalycis Herba (MHH) from different districts in China were analyzed by multidimensional data. High-performance liquid chromatography (HPLC) was used for the analysis of non-volatile indicator components, and a Heracles NEO ultra-fast gas phase electronic nose (UF-GC-e-nose) was used for the analysis of volatiles. In addition, computer vision techniques were used to determine the color and texture characteristics of samples. Besides the distinctive volatile components in different growing areas, 17 characteristic factors were screened by multivariate statistical analysis to identify the geographical origin of MHH. Moreover, the Whale Optimization Algorithm-Deep Belief Network (WOA-DBN) classification algorithm was developed and optimized in tracing the geographical producing area of MHH. The accuracy was significantly improved in comparison with regular discriminant analysis methods, such as principal component analysis (PCA) or partial least squares discriminant analysis (PLS-DA). This work provides a reference for food geographical origin traceability and quality assessment by constructing intelligent algorithms based on multidimensional data fusion.